How to keep up with protein design

I read a lot of protein design news, because we make a free monthly roundup on it. It's one of the most interesting fields I know, and one of the hardest to keep up with. This is what I've learned about doing it without it taking over your week.

The short answer: pick a small number of sources you trust, check them at a regular time, and write down why something mattered when you save it. Most of this guide is about why that's harder in protein design than it sounds, and the ways people usually go about it.

Why is protein design hard to keep up with?

Mostly because it isn't one field. To a protein engineer it might mean a new binder or enzyme, to a structural biologist it might mean cryo-EM or a new way to model dynamics, and to a computational biologist it might mean a protein language model, a structure predictor or a new benchmark. Biotech companies add another stream of therapeutic candidates, partnerships and tools.

Look at this month's stories. Guo and colleagues designed proteins that switch between specified conformations, with four NMR structures validating the designs. DreamFold reported folding-pathway generation up to roughly 30,000 times faster than molecular dynamics. A study found that changing AlphaFold3's multiple-sequence alignments improved alternative-conformation recovery on a 107-protein benchmark. RLXF used experimental feedback to improve protein language models across five protein families. Fold'EM proposed inferring structures directly from cryo-EM particles.

The stories also need different levels of caution. SynthIDBio kept its hidden signature while preserving binding affinity in tests, but redesigning a marked protein could remove it while keeping function. Fold'EM is a preprint claim, and DreamFold's trajectories are model-generated. A promising computational result, a lab-tested design and a therapeutic candidate entering screening are different stages, even when they appear in the same roundup.

How do people usually keep up?

Most people use some mix of these:

• Newsletters. Good for a weekly overview, because somebody else has done the selecting. The catch is that their selection may favour model releases or company news over the experimental detail you need.
• Journal alerts and preprint feeds. The closest you get to new methods and results. They're also noisy, and it takes time to tell a real validation from an early computational claim.
• Podcasts. Easy to fit into a commute, and good for context across design, structure prediction and biotech. Less good for remembering a sequence, benchmark or caveat later.
• Conferences and meetings. The best way to hear what people are working on before it's published, and to ask how a result was tested. They only happen a few times a year, so they can't be your only source.
• Colleagues and communities. Often how useful preprints, datasets and tools reach you. It works until the person who normally sends them is busy, or everyone is following the same exciting result.

None of them covers everything, so most people end up with a patchwork.

What tends to work

• Keep the list short. Three or four sources you trust are better than fifteen you feel guilty about. If one hasn't given you anything useful in a month, drop it.
• Pick a time. Fifteen minutes on the same day each week works better than catching up whenever you can. A monthly check can work for slower-moving topics, but fast-moving model and preprint work will pile up.
• Write one line when you save something. Why it matters to your work, in a sentence. Say whether it was computational, tested in the lab, or shown to work in cells or animals. A bare link rarely gets looked at again.
• Go back to the source for anything you'll act on. A summary tells you what something is about. For the exact success rate, structure, benchmark or caveat, read the original.

That's most of it. It takes discipline, and protein design moves between papers, preprints, tools and company announcements quickly enough to make a regular routine worthwhile.

One more option

This is why I built Iris. You describe your work in your own words, something like "de novo binder design for therapeutic proteins using diffusion models", and every month it reads across the field and writes you a briefing on what changed for you, with every source linked so you can check anything yourself. It can separate a computational result from experimental validation and keep the less relevant parts out of the way.

It's one option, and a good routine with a couple of newsletters works well too. If you'd like audio, our free monthly podcast, Keeping up with Protein Design, is about five minutes on the month's biggest stories, with the written report and every source underneath. When you'd like one built around your own work, you can set up your own briefing.

If you work in protein design or structural biology and think I've missed something, I'd like to hear about it: joe@irisbriefings.com.